Worked examples: fictional scenarios showing how I approach real infrastructure decisions - the reasoning before the implementation, not a list of services.
The companies and hardware are invented. The reasoning is the same thing you’d get on a live project.
How I’d turn an un-inventoried Heroku estate into a versioned, queryable snapshot without putting secrets in it.
A worked example of the reasoning behind an HA virtualization cluster: what “highly available” actually has to mean, where it breaks, and how I prove the design before it carries anything real.